Women’s employment, income and divorce in West Germany: a causal approach

Journal for Labour Market Research

Table 8 Balance quality for raw, matched and weighted sample

	Number		Normalized difference		Kolmogorov–Smirnov
	Treated	Control	Max ND	Mean ND	Max KS	Mean KS
Main sample
Raw sample	410	1436	0.988	0.125	0.186	0.061
Logit model and kernel matching	410	956	0.078	0.012	0.066	0.020
GBM model and odd weighting	410	1104	0.229	0.037	0.051	0.026
Group A—women not regularly employed in t₋₂₁₈₉ to t₋₇₃₀
Raw sample	142	642	0.510	0.135	0.163	0.044
Logit model and kernel matching	142	430	0.083	0.036	0.094	0.023
GBM model and odd weighting	142	428	0.203	0.065	0.088	0.024
Group B—women with strong labor market attachment (regular employment) in t₋₂₁₈₉ to t₋₇₃₀
Raw sample	112	293	0.691	0.111	0.151	0.041
Logit model and kernel matching	112	177	0.079	0.031	0.113	0.033
GBM model and odd weighting	112	160	0.154	0.045	0.099	0.029

Numbers in column 1 and 2 refer to the effective sample size. Numbers for the control group are based on either the kernel weights or the odd weights. Briefly, the effective sample size gives an estimate of the number of controls that are comparable to the treatment group. For an introduction, see for example Ridgeway et al. (2015: 9)